Getting Your Project Funded

The Heilmeier Catechism

The Defense Advanced Research Projects Agency (DARPA) funds a large portion of the fundamental research in the United States. Fundamental research is the sort of research that aims to advance scientific knowledge, and that may or may not lead to a big financial payoff or technological breakthrough. Researchers will submit proposals to the agency to investigate specific areas of interest, and DARPA will select the most promising. Unfortunately, “promising” is not a definable metric, and it can be easy for bias to creep in. George Heilmeier, a former DARPA director, developed the Heilmeier catechism, or a framework for evaluating research proposals. It is also useful when developing programs, as it forces you to think through all aspects of it. Below I will list of the tenets of the framework and discuss how I and others who work in research fulfill the requirements.

• What are you trying to do? Articulate your objectives using absolutely no jargon.

  • The further you get into the technial world, the harder it can be to know what the general public does and does not understand. It is an art that takes practice, but the government does provide “plain-language” guidelines to help.

• How is it done today, and what are the limits of current practice?

  • This is essential the same as a literature review. Google Scholar contains troves of peer-reviewed articles that should be cited. Newer fields such as machine learning might also get away with looking at blogposts from Big Tech players like Google, or prominent bloggers on websites like Towards Data Science.

• What is new in your approach and why do you think it will be successful?

  • Hopefully you already know why your approach is new. Literature reviews can help bolster your case for why your idea will be successful - if a method works in one domain, you can use it as a reason for it working in a different domain. Small examples can also go a long way.

• Who cares? If you are successful, what difference will it make?

  • This answer should be tailored to fit your audience. If you have an idea for a new treatment method for PTSD, and you are presenting to the US Army, you should talk about how it will benefit combat veterans and not women suffering from domestic abuse. If you’re proposing to DARPA, the answer should always include some part of the US government.

• What are the risks?

  • This can be a tricky question to answer, especially if you are new to proposals. A good approach is to gather a group of experts in related areas, then walk through your proposal and see what they have to say. As a software person, I should always run my new spacecraft algorithm ideas by electrical engineers and mechnical engineers who build the spacecraft the algorithm would run on. They have a different perspective and can offer expertise I lack.

• How much will it cost?

  • Expert guidance will be needed to determine the number of hours required by each technical staff member, and most companies will have an internal system that can can convert that into a cost that includes non-technical staff, such as IT workers and administrative assistants. Required equipment should also be factored in.

• How long will it take?

  • Traditional engineering tools such as Gantt charts or more software-focused websites like Jira can be used to estimate this, but input is needed from upper management to understand the company’s staff availability. It doesn’t matter if something can be accomplished in 10 weeks if there are no employees available to work on it. Extra time for illness, vacation, and other life events are also built into this.

• What are the mid-term and final “exams” to check for success?

• In my experience, the agency that is soliticing the proposal will often provide this. But if they don’t, developing them should be easy if your idea has been fully thought through by answering the other questions. The Gantt and Jira charts require listing out intermediate steps, which should lead to the mid-term exams. The final exam should simply be a test to see if your idea works!